Netskope Announces AI Guardrails Solution Leveraging Google Cloud TPU
Netskope has expanded its partnership with Google Cloud to launch 'Netskope One AI Guardrails,' optimizing security and performance for AI workflows by leveraging TPU technology. This solution enables real-time protection for autonomous AI agents, allowing enterprises to scale AI adoption securely.
📋 Article Processing Timeline
- 📰 Published: May 27, 2026 at 00:10
- 🔍 Collected: May 26, 2026 at 15:31
- 🤖 AI Analyzed: May 27, 2026 at 06:19 (14h 47m after Collected)
Netskope (NASDAQ: NTSK), a global leader in secure access and networking for the cloud and AI era, announced on April 22, 2026, an expansion of its partnership with Google Cloud to provide performance and security in AI workflows. With this new solution, enterprises can leverage Netskope One AI Guardrails to deploy high-performance generative AI and autonomous agent workflows on Google Cloud at scale. By utilizing Netskope’s capabilities, this new solution enhances the safety of these AI workflows and the data they handle.
Global enterprise investment in AI is expected to exceed $867.3 billion by 2029. As organizations transition from simple chatbots to autonomous AI agents that perform tasks independently, native, real-time protection is becoming essential. As part of Netskope One AI Security, Netskope One AI Guardrails protects data, thwarts AI-specific threats, and enables responsible AI use through content moderation capabilities.
By leveraging Google Cloud Tensor Processing Units (TPUs), Netskope delivers industry-leading cost efficiency and high-throughput inference infrastructure, providing security for even the fastest AI workloads with low latency. Using this solution, safety checks can be executed rapidly on optimized TPU instances. Netskope achieves extremely low latency by performing in-line processing aligned with the model’s native processing speed without compromising performance. Furthermore, thanks to real-time moderation via Google Cloud’s Vertex AI platform, Netskope One AI Guardrails utilizes the dedicated accelerator architecture supporting Google Cloud’s AI services.
Netskope One functions to protect autonomous agents, verifying every step against corporate policies in real-time as agents invoke tools, communicate with MCP servers, or execute autonomous tasks. The solution detects and blocks unintended recursive loops or malicious "ghost commands" that could lead to resource exhaustion or unauthorized system changes on Google Cloud. Additionally, it conducts real-time safety checks across categories including hate speech, discrimination, illegal acts, self-harm, piracy, copyright infringement, and requests for sensitive information.
While traditional security measures focus on data-at-rest, Netskope actively identifies and blocks AI-specific threats such as prompt injection and jailbreaks before the model's integrity is compromised. As the use of MCP, APIs, and other tools by agents expands, Netskope inspects the data exchanged between agents and tools, preventing indirect prompt injection and unauthorized execution. All detection results are mapped to industry-standard frameworks like MITRE ATLAS™ and the OWASP Top 10 for LLMs, providing security teams with a clear, auditable overview of risk.
Because Netskope's solution is deployed directly within the customer's Google Cloud instance, sensitive prompts and responses are scanned locally, ensuring data protection without leaving the managed tenant. By applying consistent policies across all AI communications, customers can more easily meet stringent requirements such as GDPR, HIPAA, and the EU AI Act.
Vineet Bhan, Director of Security and Identity Partnerships at Google, stated, "The collaboration between Netskope and Google Cloud maximizes the strengths of Google Cloud's AI innovation and Netskope's enterprise security. Netskope One AI Guardrails works with our Vertex AI and TPUs to enable the secure deployment of both generative AI and autonomous agents." Sanjay Beri, CEO and co-founder of Netskope, commented, "As a global leader in AI-ready security and networking, Netskope provides essential capabilities to deploy the latest AI workflows securely and efficiently. We are proud to work with Google Cloud to help our customers pursue their AI goals with confidence." Netskope products are available via the Google Cloud Marketplace.
Global enterprise investment in AI is expected to exceed $867.3 billion by 2029. As organizations transition from simple chatbots to autonomous AI agents that perform tasks independently, native, real-time protection is becoming essential. As part of Netskope One AI Security, Netskope One AI Guardrails protects data, thwarts AI-specific threats, and enables responsible AI use through content moderation capabilities.
By leveraging Google Cloud Tensor Processing Units (TPUs), Netskope delivers industry-leading cost efficiency and high-throughput inference infrastructure, providing security for even the fastest AI workloads with low latency. Using this solution, safety checks can be executed rapidly on optimized TPU instances. Netskope achieves extremely low latency by performing in-line processing aligned with the model’s native processing speed without compromising performance. Furthermore, thanks to real-time moderation via Google Cloud’s Vertex AI platform, Netskope One AI Guardrails utilizes the dedicated accelerator architecture supporting Google Cloud’s AI services.
Netskope One functions to protect autonomous agents, verifying every step against corporate policies in real-time as agents invoke tools, communicate with MCP servers, or execute autonomous tasks. The solution detects and blocks unintended recursive loops or malicious "ghost commands" that could lead to resource exhaustion or unauthorized system changes on Google Cloud. Additionally, it conducts real-time safety checks across categories including hate speech, discrimination, illegal acts, self-harm, piracy, copyright infringement, and requests for sensitive information.
While traditional security measures focus on data-at-rest, Netskope actively identifies and blocks AI-specific threats such as prompt injection and jailbreaks before the model's integrity is compromised. As the use of MCP, APIs, and other tools by agents expands, Netskope inspects the data exchanged between agents and tools, preventing indirect prompt injection and unauthorized execution. All detection results are mapped to industry-standard frameworks like MITRE ATLAS™ and the OWASP Top 10 for LLMs, providing security teams with a clear, auditable overview of risk.
Because Netskope's solution is deployed directly within the customer's Google Cloud instance, sensitive prompts and responses are scanned locally, ensuring data protection without leaving the managed tenant. By applying consistent policies across all AI communications, customers can more easily meet stringent requirements such as GDPR, HIPAA, and the EU AI Act.
Vineet Bhan, Director of Security and Identity Partnerships at Google, stated, "The collaboration between Netskope and Google Cloud maximizes the strengths of Google Cloud's AI innovation and Netskope's enterprise security. Netskope One AI Guardrails works with our Vertex AI and TPUs to enable the secure deployment of both generative AI and autonomous agents." Sanjay Beri, CEO and co-founder of Netskope, commented, "As a global leader in AI-ready security and networking, Netskope provides essential capabilities to deploy the latest AI workflows securely and efficiently. We are proud to work with Google Cloud to help our customers pursue their AI goals with confidence." Netskope products are available via the Google Cloud Marketplace.
FAQ
What is the key advantage of this product?
It uses in-line processing with TPU to identify and block advanced threats like prompt injection instantly, without compromising AI inference speed.
What kind of risks can it mitigate?
It prevents risks such as unintended recursive loops, malicious ghost commands, hate speech, and leaks of sensitive information in real-time.
How does it differ from traditional security measures?
Unlike traditional data-at-rest protection, it verifies and protects the 'behavior' of AI as it processes data in real-time.